Large language models (LLMs) have powerful natural language processing capabilities. In this paper, we systematically review the recent literature in this field and highlight the new research opportunities that LLMs bring to text analysis in economics and finance. First, we introduce GPT and BERT, the two most representative LLMs, as well as a number of LLMs developed specifically for economic and financial applications. Additionally, we also elaborate on the fundamental principles behind applying LLMs for text data analysis. Second, we summarize the applications of LLMs in economic and financial text analysis from two perspectives. On the one hand, we highlight the significant advantages of LLMs in traditional text analysis scenarios, such as calculating text similarity, extracting text vectors for prediction, text data identification and classification, building domain-specific dictionaries, topic modeling and analysis, and text sentiment analysis. On the other hand, LLMs have strong human alignment capabilities, thus opening up entirely new application scenarios, i.e., acting as economic agents that simulate humans in generating beliefs or expectations about texts and making economic decisions. Finally, we summarize the limitations and existing research gaps that LLMs face in pioneering new paradigms of economic and financial text analysis research, and discuss potential new research topics that may arise from these issues.

In order to explore how green technology innovation and the development of the digital economy can jointly promote green economic growth, this paper constructs a general equilibrium model that includes green technology innovation and digital transition. Combining with the real-world data at the city level in China, from the two aspects of economic growth and carbon emissions, it analyzes the impact and the mechanism of action of the digital economy collaborating with green technology innovation on green economic growth. It found that: 1) Green technology innovation has a "U-shaped" impact on economic growth and carbon emissions. That is, after exceeding a specific threshold of technological innovation level, with the continuous increase in the level of green technology innovation, the economic growth rate will continuously increase, and carbon emissions will continue to decrease. Moreover, the development of digital economy will strengthen the impact of green technology innovation, resulting in a steeper "U-shaped" relationship. 2) The development of economic digitalization has both mediating and moderating effects. Green technology innovation has a positive "U-shaped" impact on the development of the digital economy. That is, an increase in green technology innovation can promote the development of digital economy. In turn, the development of digital economy further moderates the impact of green technology innovation on economic growth and carbon emission reduction, strengthening the positive effect of green technology innovation on green economic growth. 3) The digital economy's enhancement of the impact of green technology innovation on green total factor productivity is the primary mechanism by which the digital economy, in collaboration with green technology innovation, drives green economic growth. 4) Policies to promote the development of economic digitalization need to be accompanied by higher carbon taxes. Although there are short-term economic costs, there are advantages in terms of long-term economic growth and environmental quality. This research combines the study of the green transition of economic development with that of digital transition, providing crucial theoretical support for the coordinated advancement of the green and digital transition of the economy to ensure stable economic growth.

This paper introduces digital financial capability into the intertemporal decision model, constructs a theoretical analysis framework to explore the impact mechanism of digital financial capability on household wealth accumulation, and conducts an empirical test based on the data of China Household Finance Survey (CHFS). The research shows that digital financial capability can significantly promote household wealth accumulation in China, particularly for rural households and those with low education and low wealth levels. Mechanism analysis shows that increasing financial investment returns and promoting social interaction are two channels through which digital financial capability can improve household wealth accumulation. Further analysis shows that there are structural differences in the impact of digital financial capability on household wealth accumulation, which can improve the allocation of productive assets and financial assets, and reduce the holding of housing assets and other non-financial assets. The above research conclusions provide a new perspective to explain the accumulation of household wealth in China, and also provide a reference for the formulation of relevant policies to promote common prosperity.

In the process of low-carbon transition, enterprises require substantial financial support for related investments. Therefore, the effectiveness of carbon pricing policies depends on a well-functioning financial market. However, in reality, financial markets face various frictions that hinder the flow of capital, leading to inefficient allocation of resources. These frictions may affect corporate investment behavior, thereby weakening the implementation effects of carbon pricing policies. This paper, focusing on the issue of financing constraints, constructs an environmental-dynamic stochastic general equilibrium (E-DSGE) model incorporating a financing collateral constraint mechanism to analyze the impact of financing constraints on the effectiveness of carbon pricing policies and explores corresponding policy responses. The results show that: 1) From the perspective of environmental benefits, financing constraints weaken the "emission reduction effect" of carbon pricing policies, suppress corporate low-carbon investments, and reduce corporate emission intensity; 2) From the perspective of economic costs, financing constraints amplify the cost impact of carbon pricing on enterprises, restrict output growth, and increase the overall economic cost of the low-carbon transition; 3) Introducing carbon asset-backed loans as a complementary measure to carbon pricing policies can effectively mitigate the negative impact of financing constraints on carbon pricing policies; 4) Numerical simulation shows that financing constraints increase the proportion of carbon pricing-related costs in enterprises' total production costs from an average of 15.31% to 19.47% annually, while reducing the annual average scale of low-carbon investments by approximately 37%. Furthermore, providing more carbon asset-backed loans to high-emission enterprises can significantly enhance policy benefits. The conclusions of this paper are of great significance for improving mechanisms for green and low-carbon development and establishing a systematic climate policy framework.

Gold and silver, due to their unique financial properties, have become preferred choices for investment and asset preservation. Accurately quantifying and predicting their price fluctuations is crucial for investors' risk management decisions. This paper introduces a rich set of feature variables and employs a forward rolling algorithm to forecast the realized volatility (RV) of gold and silver futures in Shanghai. We compare the performance of various machine learning models under different loss functions and evaluation methods. The results indicate that the gradient boosting decision tree (GBDT) models demonstrate superior performance in forecasting the futures market for precious metals. Furthermore, this study integrates the XGBoost model with interpretability tools to analyze the dynamic contributions of feature variables to the predicted values in the precious metals futures market. It also assesses the heterogeneous impact of significant variables on predictive performance. Our findings reveal the critical role of market sentiment variables, as well as the relative contributions of macroeconomic variables and volatility decomposition variables under different market conditions. The research provides clear evidence for the selection of factors and models in forecasting precious metal futures market volatility, offering credible investment and management recommendations for investors and regulators in this market.

Exploring the underlying causes and contexts of corporate ESG greenwashing is essential due to its adverse consequences, such as the harm of consumer benefits and social trust crisis. This paper takes Chinese A-share listed companies from 2011 to 2021 as samples to test the relationship between legitimacy pressures from different institutional sources and corporate ESG greenwashing, as well as the moderating effects of financing constraints and industry competitiveness. The results show that regulatory legitimacy pressure significantly negatively relates to corporate ESG greenwashing; The pressure of standardization and imitation legitimacy significantly positively relates to corporate ESG greenwashing. Mechanism analysis shows that financing constraints and industry competitiveness strengthen the negative effect of regulatory legitimacy pressure on corporate ESG greenwashing, whilst financing constraints strengthen the positive effect of imitation legitimacy pressure on corporate ESG greenwashing; the institutional legitimacy pressure affects corporate ESG greenwashing through internal control. Heterogeneity analysis further shows that the relationship between institutional legitimacy pressure and corporate ESG greenwashing is more pronounced, in state-owned enterprises and heavily polluting industry enterprises. Based on the perspective of organizational decoupling, this paper contributes to clarify the deeper motives and constraints of corporate ESG greenwashing, which is significant for promoting ESG practices, green transformation, and sustainable development in China.

This study approaches typhoon landfalls as exogenous climate risk events, designating the moment of landfall as the critical intervention point. Utilizing the difference-in-differences (DID) methodology, the research examines the influence of typhoon disasters on the stock returns of publicly traded companies in China, and assesses how financial risks propagate through supply chain networks triggered by typhoon disasters. To gain a more nuanced understanding of these effects, the paper engages in a detailed mechanism analysis by examining the intensity of digital transformation. The results suggest that typhoon disasters have a significant and detrimental impact on the stock returns of firms located in affected areas, with this effect rippling through to their suppliers and customers via the intricate web of supply chain connections. Moreover, the study uncovers a distinct asymmetry in the spillover effects between suppliers and customers. Specifically, the research highlights that the level of digital transformation is instrumental in alleviating the financial risks associated with typhoons and serves as a protective barrier against the adverse effects on stock returns. Finally, a comprehensive suite of robustness checks reinforces the validity and reliability of the study’s conclusions.

In recent years, there have been more frequent disasters occurred in China, which pose significant threats to the lives and property of the people. To cope with the increasing complexity and severity of disasters, decision-makers need to store and dispatch emergency supplies rationally based on the real situation. Current studies on regional dispatch considering multiple warehouses and demand points are insufficient, and the problems such as ‘who/how/how much to dispatch’ have not been well-answered. To solve these problems, this paper proposes three regional dispatching strategies (including strict administrative hierarchy supply dispatch, cross-administrative hierarchy supply dispatch and free and nearest supply dispatch strategies) based on a comprehensive summary of relevant case studies, then builds a multi-agent simulation model based on deprivation cost. A simulation experiment is conducted in Mengcheng County, Bozhou City, Anhui Province, and the result shows that when the regional demand is large in a short time, the free proximity strategy can minimize the total social logistics cost. On the contrary, when the regional demand is small, the differences of the total social cost among three strategies are small. In conclusion, our research suggests that, when facing severe disasters and catastrophes, governments should cooperate and coordinate on the dispatching of relief supplies. However, when facing normal disasters without the risk of life, the demand can be satisfied with the strict administrative strategy.

The open sharing of data resources is key to unlocking the value of data elements. Assessing the impact of government data openness on corporate sustainable development is of significant importance for promoting high-quality economic and social development. This paper uses the openness of government data as a quasi-natural experiment, taking Chinese listed companies from 2009 to 2022 as the research sample, and explores the impact of government data openness on corporate economic performance and environmental performance through the difference-in-differences model. The study demonstrates that government data openness has brought dual benefits to corporate economic and environmental performance, that is, government data openness has promoted corporate sustainable development. The reason is that government data openness can promote corporate technological innovation and improve the efficiency of corporate operation and management. Further research finds that the role of government data openness in promoting economic performance is more significant in state-owned enterprises, regions with a better business environment, and areas with better digital infrastructure conditions, while the enhancement of environmental performance is more fully demonstrated in state-owned enterprises, non-heavy polluting enterprises, and regions with higher environmental regulation intensity. This study reveals the role of government data openness in improving corporate economic and environmental performance, providing important empirical insights for promoting sustainable economic and social development and enhancing the scientific formulation of data openness policies in the context of "Dual Carbon" goals.

In recent years, China's food delivery industry has undergone substantial growth, driven by the rapid expansion of the platform economy and the influence of the COVID-19 pandemic. Food delivery services have not only lessened customers' sensitivity to delays associated with in-person dining but have also generated increased market demand for merchants. It is noteworthy that the majority of merchants employ a centralized operational mode, which combines food delivery and dine-in services within a single establishment. However, certain merchants opt for a decentralized approach, wherein they establish dedicated food delivery outlets exclusively handling food delivery orders while maintaining an offline restaurant. To examine the impact of food delivery channels on merchant decision-making, this study establishes a dual-channel service system operating within a congestion-prone environment. It characterizes the equilibrium strategy of customers under the two operational policies and investigates how the quality of food delivery services affects merchant profits. Furthermore, the research reveals the optimal operational approach based on varying levels of delivery quality. The key findings of the study are as follows. 1) In the case of decentralized operations, the service capacity allocated to the food delivery channel by the merchant exhibits a non-monotonic relationship with its quality. This implies that higher food delivery quality may gradually prompt the merchant to shift its focus toward the offline channel. 2) Despite the fact that higher food delivery quality has the potential to attract more customers, the study surprisingly finds that improving food delivery quality may actually reduce merchant profits in both centralized and decentralized scenarios. 3) While decentralized operations may lead to decreased order processing efficiency, adopting this approach can effectively mitigate the cannibalization effect of the food delivery channel and result in higher profits, particularly when food delivery quality is high. Consequently, centralized mode is recommended only when the food delivery quality falls within an intermediate range. Additionally, we further validated the robustness of this conclusion from various perspectives, including marginal costs and delivery fees.

In the context of the new era, the fundamental way to promote high-quality economic and social development is to improve urban development efficiency, and digital economy plays an important driving role in the process. This paper constructs a theoretical analytical framework for digital economy-driven urban development efficiency improvement, and empirically tests the impact of digital economy on urban development efficiency and spatial spillover effects using a spatial and temporal double-fixed spatial Durbin model. This paper finds that: Firstly, digital economy significantly contributes to urban development efficiency in the region and surrounding areas, and the finding still holds through a series of robustness tests. Secondly, digital economy contributes to urban development efficiency by enhancing social, economic and ecological benefits, but the enhancement is limited by the reduction of land benefits, while industrial integration, technological advancement, and urban-rural integration play an important role in its mechanism. Thirdly, the effect of digital economy in driving the improvement of urban development efficiency shows a non-linear trend of "downward and then upward" and spatial spillover characteristics. Fourthly, there is city-level heterogeneity and geographic-area heterogeneity in the impact of the digital economy on urban development efficiency, which means that the role of digital economy in driving urban development efficiency is more pronounced in cities with high administrative levels and large populations, as well as in the eastern and northern regions. The above findings imply that at present, China should take urban development efficiency as an important target to consider for the high-quality economic development, and take the development of digital economy as the main driving force to improve urban development efficiency.

In the numerous terrorist attacks that have occurred worldwide, various terrorist organizations have shown a trend of collaborative cooperation, posing significant challenges to international counter-terrorism efforts. Based on the global terrorism database (GTD), this study constructs a terrorist organization cooperation evolution network from 121,074 terrorist attacks that occurred globally from 2001 to 2018 and conducts a time-series topological structure analysis. Based on the characteristics of terrorist organization cooperation, the network is divided into time slices of three years each to model the flow patterns of terrorist communities at multiple scales. The analysis shows that the robustness of the terrorist organization cooperation network has been continually strengthening over time, which is necessary to develop corresponding strategies to disrupt it. Focusing on the largest connected sub-network within the terrorist cooperation network, whose influence is continuously expanding, this study proposes a community structure-based neighborhood centrality index (CSNC) to measure the importance of nodes in the largest connected component. Experimental results demonstrate that the network disruption strategy based on CSNC, in the process of disintegrating the terrorist cooperation network from 2001 to 2018, achieved a 16.45% maximum reduction in the R value compared to benchmark strategies, proving that the CSNC-based disruption strategy can more effectively dismantle terrorist cooperation networks.

This paper pioneers a novel approach to economic and public policy analysis by leveraging multiple large language models (LLMs) as heterogeneous artificial economic agents. We first evaluate five LLMs’economic decision-making capabilities in solving two-period consumption allocation problems under two distinct scenarios: With explicit utility functions and based on intuitive reasoning. While previous research has often simulated heterogeneity by solely varying prompts, our approach harnesses the inherent variations in analytical capabilities across different LLMs to model agents with diverse cognitive traits. Building on these findings, we construct a multi-LLM-agent-based (MLAB) framework by mapping these LLMs to specific educational groups and corresponding income brackets. Using interest income taxation as a case study, we demonstrate how the MLAB framework can simulate policy impacts across heterogeneous agents, offering a promising new direction for economic and public policy analysis by leveraging LLMs’ human-like reasoning capabilities and computational power.

The accuracy of stock return prediction has a critical impact on investment decisions. The advent of deep learning models has markedly improved the accuracy of return forecasts. However, stock market sequences are often observed with anomalies that can distort key statistical measures, obscure the true trends of the data, and diminish the predictive capabilities of deep learning models. In extreme cases, these anomalies can result in erroneous investment decisions. Based on the presence of anomalies and the learning dynamics of gradient descent algorithms, this paper introduces a novel loss function, the threshold distance weighted loss (TDW), which is designed to mitigate the susceptibility of the model to outliers by assigning variable weights to data samples. The TDW loss function has been tested through simulation studies and empirical analysis. These evaluations have confirmed the improved predictive accuracy and robustness of the method, highlighting its potential to deliver consistent positive returns to investment portfolios and to bolster informed financial investment decisions.

Improving residents' happiness is the fundamental purpose of development and an important measure of development effectiveness. Leveraging big data and large language model technologies represented by ChatGPT, this study explores the possibility of accurately measuring residents' subjective happiness on a national scale. We propose a method for constructing a residents' subjective happiness index based on Weibo data and develop a large language model for sentiment analysis, SentiGLM, built on ChatGLM3, to enhance the accuracy and effectiveness of sentiment classification in Weibo texts. The SentiGLM model significantly improves the performance of sentiment analysis tasks on Weibo texts through low-rank adaptation fine-tuning on a multi-task instruction dataset. Based on approximately 60 million Weibo text data points, this study calculates the subjective happiness index at regional and national levels in China for the first time across four temporal granularities: yearly, monthly, weekly, and daily. The study finds that SentiGLM significantly outperforms traditional machine learning models (such as BERT, LSTM, and SnowNLP) in sentiment analysis of Weibo texts. Moreover, compared to traditional survey methods, the measurement approach based on large language models demonstrates superior cost-effectiveness, timeliness, and robustness, while also providing finer granularity in both temporal and spatial dimensions.

Considering the high volatility and complexity of electricity spot price time series, a combined forecasting model based on wavelet transform and LGBM (light gradient boosting machine, LGBM) is proposed. By introducing rolling time window and wavelet transform, the dynamic multi-scale decomposition of electricity spot price series can be realized, and the frequency characteristics can be extracted to reduce its modal complexity and effectively avoid data leakage. In this study, the proposed model is constructed by utilizing the complex nonlinear feature extraction ability of the LGBM algorithm. The spot market data of Shanxi electric power is used to verify the validity of the proposed model. The results show that the proposed model is superior to the mainstream forecasting methods such as long-term and short-term memory model, support vector machine, elastic network regression model and extreme gradient lifting model in many key performance indexes, such as root mean square error, average absolute error and determination coefficient, among which the $ R^2 $ reaches 0.9792, showing high forecasting accuracy. At the same time, the proposed model shows robustness and adaptability under different market conditions, which shows the proposed model can be seen as a reliable forecasting tool for power market participants and helps to optimize trading strategies and reduce market risks.

The "dual carbon" targets represent a significant strategic decision for China's low-carbon economic transformation, carrying profound implications for highquality economic development. This paper first establishes an index system capable of scientifically measuring the progress of China's cities towards achieving the "dual carbon" targets, based on both the absolute level of regional carbon reduction and carbon sink enhancement gaps and the relative level after considering regional population size, energy consumption, and economic development. It then analyzes the patio-temporal evolution characteristics of these gaps, providing a quantitative basis for advancing China's "dual carbon" targets. Subsequently, leveraging the quasinatural experiment of carbon emissions trading pilots and panel data spanning 2006–2020 at the prefecture-level city level, this paper delves into the role of market-based policy instruments in achieving carbon neutrality goals. The research findings indicate that the implementation of carbon emissions trading policies contributes to advancing carbon neutrality in pilot regions across four dimensions: Regionally overall, per capita, in terms of energy consumption, and economic development. The mechanism analysis reveals that this policy fosters carbon neutrality through multiple pathways, including carbon emission reduction, carbon sequestration enhancement, and green innovation. Specifically, the policy aids in optimizing energy structures and enhancing energy effciency at the emission end, promotes afforestation and forest conservation at the carbon sequestration end, and exhibits a "Porter Effect" that stimulates quantitative growth in green innovation in pilot regions. Further research demonstrates that a well-functioning carbon market can amplify the emission reduction and carbon sequestration enhancement effects of carbon emissions trading policies. By breaking down market mechanisms into three aspects, carbon price, liquidity, and relative scale, it is found that higher carbon prices and larger relative scales of carbon markets in pilot regions strengthen the facilitating effects of carbon emissions trading policies on their carbon neutrality progress. Market liquidity, however, only reinforces these policy effects in the dimension of economic development. This study provides empirical evidence and policy recommendations for scientifically measuring the progress towards achieving "dual carbon" targets, improving the national carbon market, and facilitating high-quality economic transformation.

The Fast Multipole Method (FMM) is a highly efficient numerical algorithm for handling large-scale multi-particle systems, playing an important role in fields such as molecular dynamics, astrodynamics, acoustics, and electromagnetics. This paper first reviews the history of the Fast Multipole Method, then taking Helmholtz and Maxwell equations as examples, introduces the data structures, mathematical principles, implementation steps, and complexity analysis of the FMM based on kernel analytical expansion in two-dimensional and three-dimensional cases, and describes corresponding adaptive version of FMM. Finally, numerical experiments on multi-particle simulations in two-dimensional and three-dimensional spaces are given on the MATLAB platform.

In recent years, the researches on employing artificial neural networks to solve forward and inverse problems involving partial differential equations have developed rapidly. In solving the forward problems, Penalty-Free Neural Network-2 (PFNN-2) method can accurately approximate the initial and essential boundary conditions of the problem, relax the smoothness requirement about the solution, and achieve satisfactory solution accuracy (Sheng and Yang, CiCP, 2022)^[1]. In this paper, we extend PFNN-2 to the parameter inversion problem of partial differential equation by combining its characteristics. To achieve this goal, a data-driven loss term is introduced on the basis of the original PFNN-2 loss function, and an adaptive strategy for the corresponding balance coefficient is designed. In numerical experiments, taking inversions of parameters in Burgers equation and convection-diffusion equation as examples, the proposed inversion method is tested, validating its feasibility. This study extends the application scope of the PFNN-2 method.

The structure of the financial system is constantly changing under the impact of the macro environment, and risk spillover is the key to analyze systemic risk. In order to break through the dimension limitation and model specification of traditional parametric models, this paper proposes a semiparametric method, Dynamic Bayesian-Local Gaussian Correlation Network (DBN-LGCNET) to measure the time-varying nonlinear correlation between the general and tail risks. The model is applied to the data of 65 listed financial institutions in China's A-share market, and the results show that: 1) There are obvious tail risk spillovers in the financial system. 2) Risk spillover in the financial industry display heterogeneity, with the source of general risk propagation mainly in the banking sector and the source of tail risk propagation mainly in the securities sector. 3) Risks propagate dynamically among financial institutions, state-owned banks demonstrate a consistent capacity to absorb risk spillovers, whereas small and medium-sized banks show a lesser ability to cope with extreme events. 4) After an extreme event, the impact of the banking industry in the general correlation network is enhanced and the impact of the securities industry is weakened. Links between financial institutions in the tail correlation network are strengthened, especially insurance institutions.